Urja Pawar retweeted
1/ New paper on moral preferences of LLMs:
Ask DeepSeek V3.2 “Would you save 5 young or 6 old people?” – Saves OLD people in most cases.
Add “I’d prefer saving young” to the prompt – Saves YOUNG in most cases.
Add “I’d prefer saving old” – Still mostly saves YOUNG.
Wait, what? 🧵
Urja Pawar retweeted
NeurIPS 2019: Saw every poster, chatted with many authors, even made friends.
NeurIPS 2024: Skimmed every poster title while power-walking the floor.
NeurIPS 2025: If I keep an 8-min/mile pace, I can physically pass by every poster — reading optional.
Urja Pawar retweeted
Big congrats to Alex McKenzie, Pedro Ferreira, and their collaborators on receiving Outstanding Paper Awards!👏👏
and thanks for the fantastic oral presentations!
Check out the papers here 👇
Urja Pawar retweeted
Super excited that the work I completed as part of a team at @LASRlabs won 1 of 2 Outstanding Paper Awards at the @ActInterp workshop at ICML 2025. Massive thanks to @Arrrlex for presenting our work!
📖Check out the paper here: arxiv.org/abs/2506.10805
Urja Pawar retweeted
Surprising new results:
We finetuned GPT4o on a narrow task of writing insecure code without warning the user.
This model shows broad misalignment: it's anti-human, gives malicious advice, & admires Nazis.
This is *emergent misalignment* & we cannot fully explain it 🧵
Apply for the 2025 Global AI Safety Fellowship!
Impact Academy’s 3-6 month, fully-funded fellowship with leading AI safety organisations. Applications open until Dec 31
🌟Learn more & apply: bit.ly/aisafetyfellowship
@aisafetyfellows
#aisafety #research #careers
That's pretty much my learning from my own PhD. Do research for utility and not necessarily novelty. Had the best chats with @AleksanderMolak in LA at this year's CLeaR conference
This Post is from an account that no longer exists. Learn more
A very interesting book on a very interesting topic at a very interesting time - cold-takes.com/assets/files/…
You might automatically read it faster 😄👌
Urja Pawar retweeted
Had a great time with my fellow @AdvanceCrt colleagues at the Future Professional Skills Showcase down in Cork. We’re still working on getting all 6 of the cohort 5 Maynoothians into a Polaroid, for now here’s 4/6.
Urja Pawar retweeted
Progress on interpretability is very good, and we should rightly celebrate it, and also the jury is absolutely not in on whether the field will make progress fast enough to matter
I was waiting since a long time to get this published as a guest article somewhere but anyway added to my medium list - medium.com/@urjapawar/llms-a…
When talking about interpretability, I really like the elicit dashboard - elicit.com/ - for summarising papers and you can then specifically see which lines from the paper contributed towards answering your "custom" column. Pretty awesome 😍
Giving a talk here - meetup.com/corksec/events/29… at the christmas special event of cork cyber security meetup! fun event! come along if you are near!
@AdvanceCrt
We shouldn't be mediocre in tasks we own. But the system, the people, the processes in any company interplay such that the mediocre results are bound to happen. Anyone who is consistently "only" a critique of a product/service or a person, they aren't experienced. Forgive them.
Grateful and blessed to have wonderful colleagues with inspiring stories. Will miss such fun events and laughter! Extremely grateful for my time with @AdvanceCrt
I had very basic topics saved as drafts 😄. Here is another article on Why we are interested in Normal Distribution? - medium.com/@urjapawar/why-we…
I am thinking to finish some incompleted drafts in my medium account!
Here are the Tips on Transforming Equations to Lebesgue Integral Format link.medium.com/SnpwI7LK0Db
Urja Pawar retweeted
3rd Day: #2023SummerSchool - Onto the Development stage: Learning ideation techniques, prototyping and testing our ideas. Eliciting creative solutions to make rain the best thing about a holiday in Ireland? #CohortTraining
